How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Borel and Analytic Sets, Perfect Sets, and Determinacy: Examples and Counterexamples
1 · Prerequisites
- Borel and Analytic Sets, Perfect Sets, and Determinacy
- Cardinal Arithmetic, Cofinality and the Alephs
- Compactness
- Compactness in Metric Spaces
- Complete Metrizability, Čech-Completeness, and Baire Category
- Completeness, Completion, and Uniform Continuity
- Construction of the Natural Numbers
- Construction of the Real Numbers via Cauchy Sequences
- Construction of the Real Numbers via Dedekind Cuts
- Countability and Uncountability
- Countability Axioms and Cardinal Functions
- Finite Counting, Factorials and Binomial Coefficients
- Foundations of the Real Numbers for Analysis
- Metric Spaces
- Monotone Sequences, Bolzano-Weierstrass, and Cauchy Completeness
- Order, Zorn's Lemma, and the Axiom of Choice
- Ordinal Arithmetic and the First Uncountable Ordinal
- Ordinals, Cardinals, and Transfinite Recursion
- Relations, Functions, and Quotients
- Roots, Rational Powers, and Classical Inequalities
- Sequences and Limits
- Series: Convergence and the Nonnegative Tests
- Subspaces, Products, and Quotients
- Suprema and Infima
- The Cantor Set, Baire Category, and Measure Zero in ℝ
- The ZFC Axioms and the Basic Set Constructions
- Topological Spaces and Continuity
- Well-Founded Relations, Rank, and the Cumulative Hierarchy
2 · Summary
These worked examples calculate empty and singleton tree bodies, a first-move clopen game, and a well-founded code for a closed cylinder complement. A terminal-taboo counterexample explains why a winning position need not retain every child. The non-Borel refutation transfers a choice-based undetermined payoff into the real line by a continuous injection; Borel determinacy rules out Borelness of its image.
3 · Logical flowchart
4 · Definitions, theorems and proofs
None yet.
5 · Examples, counterexamples and false statements
FALSE: every Borel subset of the real line is a countable union of countable intersections of open and closed sets
Statement
Every Borel subset of belongs to the class obtained by taking countable intersections of open and closed sets and then countable unions of those intersections.
Facts & Assumptions
Given: The real line with its usual topology and the Borel classes of The countable Borel hierarchy and its limit convention. Assume the Axiom of Choice (The Axiom of Choice) for the local separation theorem below.
In ZFC, every metrizable space containing a Cantor copy has a set in (Universal Borel sets and strictness on Cantor space).
The real line is metrizable (The absolute value makes a metric space: is a metric, its open balls are the intervals , and it is unbounded).
Cantor space is compact, and continuously injects it into (Cantor and Baire sequence spaces and coordinate codings, Continuous injections of sequence spaces into the real line).
Continuous images of compact spaces are compact, a continuous bijection from a compact space to a Hausdorff space is a homeomorphism, and compact subsets of Hausdorff spaces are closed (A continuous image of a compact space is compact; a continuous real-valued map on a nonempty compact space attains a maximum and a minimum; and a continuous bijection from a compact space to a Hausdorff space is a homeomorphism, In a Hausdorff space a point and a disjoint compact set, and two disjoint compact sets, have disjoint open neighbourhoods; hence every compact subset is closed, and in a compact Hausdorff space the compact subsets are exactly the closed ones).
Refutation
Every closed subset of is a countable intersection of open sets: if , use . Points of belong to every ; a point outside has an interval disjoint from , and hence misses for sufficiently large . For use the constant empty sequence. Consequently any countable intersection of open and closed sets is a countable intersection of open sets: replace each closed factor by this explicit sequence, leave an open factor as its constant sequence, and enumerate pairs of indices by finite diagonals. Such a set is , since its complement is a countable union of closed sets. Therefore every set in the claimed class is .
The map in L3 constructs a Cantor copy in the real line. Put with its subspace topology. The real line is Hausdorff: two distinct reals have disjoint open intervals of radius around them. The subspace is therefore also Hausdorff, by intersecting these separating intervals with . As a map onto , is a continuous bijection: for open in , the preimage of is . L3 gives compactness of its domain, so L4 gives a homeomorphism . L4 also makes compact and closed in . Thus this explicit ternary-series map supplies the required homeomorphic Cantor subset without choice.
By L2 and step 1.2 the hypotheses of L1 hold for , so choose . This is a Borel set: the first three ranks are obtained from opens by complements and countable unions, operations preserving the Borel sigma-algebra. Suppose the statement were true. It would place in the displayed class and thus, by step 1.1, in , contradicting its choice. Hence the statement is false. The only choice input is L1; the conversion in step 1.1 makes no simultaneous choices of representations.
Empty and single-branch tree codes
Example
In ZF, the empty tree has empty body. For each the prefix tree has body . In particular, for , the tree has body exactly the constant-zero sequence. These bodies are closed in the cylinder topology.
Facts & Assumptions
A branch has every finite prefix in its tree, including the empty prefix; see Trees and their bodies.
Verification
Given: , as above, and the empty tree.
Every putative branch of would have to satisfy , which is false. Hence , a closed set.
Restricting to gives , so is a tree. The point lies in its body. If , its prefix of length equals the unique member of of that length. Consequently for every , giving .
For , step 1.2 gives at every coordinate, the claimed concrete calculation. If , choose one differing coordinate ; its cylinder of length excludes . Thus the complement of is open, establishing the closedness assertion. QED.
A clopen game decided by the first move
Example
On let . Player I wins by the strategy at every even-length position . The payoff is clopen.
Facts & Assumptions
Full-position strategies and their winning condition are in Gale–Stewart games and strategies.
Finite-prefix cylinders form the topology of Baire sequence space and its cylinder topology.
Verification
Given: The full natural-number tree, payoff , and the constant-zero I strategy.
Every belongs to the full tree, so is a legal strategy, including at . We have and , both open by F2. Hence is clopen.
Every branch consistent with satisfies , so belongs to . For example, if II always plays , the unique compatible branch is and its first coordinate is . The same first-coordinate calculation holds for every sequence of II moves, proving that wins. QED.
A winning taboo position can have a nonwinning child
Statement refuted
In a game with terminal taboos, a position from which a player can force a terminal taboo for the opponent need not have only such winning children. In particular, deleting all positions from which either player can force an opponent taboo need not leave a prefix-closed tree. Here infinite play does not count as successful terminal reachability for either player.
Facts & Assumptions
A terminal taboo loses for its named player, while nonterminal moves and fixed-history parity are as in Game trees with terminal taboos.
Counterexample
Given: The tree . Its only terminal node is ; declare it taboo for II. The root is I-to-move. For definiteness the infinite payoff is empty; terminal reachability ignores that payoff.
Prefixes of are the root or words with , all listed in . The only other nonempty word is , whose sole proper prefix is the root. Thus is a nonempty tree. The node is terminal and every node on the other ray has the unique child obtained by appending , verifying the asserted terminal partition.
At the root I can choose and reach the II taboo immediately. Below the child every legal continuation is forced and the only maximal continuation is the infinite sequence . No continuation from that child reaches a terminal node. Neither player can therefore force an opponent terminal taboo there, although I can do so at its parent.
The proposed deletion removes the root by step 2.1 and retains by the same step. A set retaining but omitting its empty prefix is not a tree. This is the required witness against both the child assertion and the resulting deletion rule. QED.
Remarks
This witness corrects the downward-closure assertion in the proof of Buffard–Levrel–Mayo Lemma 1 (arXiv v1). It does not refute a reduction that keeps only nodes all of whose prefixes avoid the two reachability-winning sets. It is an AI-generated counterexample and is not a dependency supplier.
Every set of reals is Borel
Statement
False assertion: every subset of is Borel.
In ZFC, take an undetermined payoff and the continuous injection supplied below. The set e[A] is a witness that the assertion fails.
Facts & Assumptions
Choice produces an undetermined natural-number game supplies A with neither player winning under AC.
Continuous injections of sequence spaces into the real line gives the continuous injection e in ZF.
Borel hierarchy exhaustion and preservation by continuous pullback makes continuous inverse images of Borel sets Borel in ZFC.
Borel games are determined determines every Borel natural-number payoff in ZFC.
Assume The Axiom of Choice.
Refutation
Given: Work in ZFC throughout this counterexample.
F1 with A1 supplies A and F2 supplies e. Suppose e[A] were Borel in . By continuity of e, F3 with A1 would make Borel in Baire space. This preimage equals A: if e(x)=e(a) for some a in A, injectivity gives x=a; conversely each a in A maps into e[A].
Then F4 with A1 would give a winning strategy for one player for payoff A, contradicting its defining property from F1. Hence e[A] is not Borel and is the required subset of the real line witnessing the failed assertion. QED.
Evaluating elementary Borel codes
Example
Fix a space with enumerated basis and . The leaf code evaluates to ; a complement above that leaf evaluates to ; an empty union evaluates to ; and a complement above an empty union evaluates to . These calculations hold in ZF.
Facts & Assumptions
Code nodes, their labels, and their well-foundedness requirement are in Well-founded Borel evaluation codes.
Existence and uniqueness of Borel-code evaluation provides unique evaluation in ZF by the three displayed rules.
Verification
Given: , and the fixed index .
Let with root label leaf. It is valid since it has no child and its empty relation is well-founded. F2 gives . For label the root complement and leaf. Then and , which is closed because is open.
On put a union label; then . On put a complement at the root and a union at . Then and .
In each two-node tree the sole child is terminal; any subset containing it has it as a minimal element, and a remaining nonempty subset is the singleton root. Thus the child relation is well-founded, and the complement nodes have exactly child zero as required. All four codes are valid even when or is empty or all of . The empty underlying tree is not one of these codes, because F1 requires a root. QED.
Sources
- D. Marker, Descriptive Set Theory, Section 2, Corollary 2.38
- M. Christ, Math 202B Lecture 1, Comment on the Borel hierarchy
- Application of Lemma 1.14(ii)
- Opening definitions of games and strategies
- Lemma 1 proof, downward-closure assertion (claim refuted by the displayed local tree)
- Exercise 6.8, p55, and Borel determinacy Theorem 6.9, p56; the continuous injection transfer is proved locally. This is a derived false-statement witness, not an attributed source claim
- Definitions 7.1–7.2, pp62–63; examples derived directly from the local labelled-code convention